Fulltext

Assessing Trustworthy AI in Times of COVID-19: Deep Learning for Predicting a Multiregional Score Conveying the Degree of Lung Compromise in COVID-19 Patients

Allahabadi H 1 , Amann J 2 , Balot I 3 , Beretta A 4 , Binkley C 5 , Bozenhard J 6 Show all authors , Bruneault F 7 , Brusseau J 8 , Candemir S 9 , Cappellini LA 10 , Chakraborty S 11 , Cherciu N 12 , Cociancig C 13 , Coffee M 14 , Ek I 15 , Espinosa-Leal L 16 , Farina D 17 , Fieux-Castagnet G 18 , Frauenfelder T 19 , Gallucci A 20 , Giuliani G 21 , Golda A 22 , van Halem I 23 , Hildt E 24 , Holm S 25 , Kararigas G 26 , Krier SA 27 , Kuhne U 28 , Lizzi F 29 , Madai VI 30 , Markus AF 31 , Masis S 32 , Mathez EW 23 , Mureddu F 33 , Neri E 34 , Osika W 35 , Ozols M 36 , Panigutti C 37 , Parent B 38 , Pratesi F 4 , Moreno-Sanchez PA 39 , Sartor G 40 , Savardi M 41 , Signoroni A 41 , Sormunen HM 42 , Spezzatti A 43 , Srivastava A 44 , Stephansen AF 45 , Theng LB 46 , Tithi JJ 47 , Tuominen J 48 , Umbrello S 49 , Vaccher F 17 , Vetter D 50 , Westerlund M 16 , Wurth R 51 , Zicari RV 16

Affiliations 

  • 1 Enterprise Intelligence DepartmentEY Netherlands 1083 HP Amsterdam The Netherlands
  • 2 Health Ethics and Policy LabDepartment of Health Sciences and TechnologyETH Zurich 8092 Zürich Switzerland
  • 3 Postgraduate Studies in Diplomacy and International RelationsCenter for Diplomatic & Strategic Studies 75015 Paris France
  • 4 Institute of Information Science and Technologies, National Research Council of Italy (CNR) 56124 Pisa Italy
  • 5 Bioethics CenterHackensack Meridian Health Edison NJ 08820 USA
  • 6 Faculty of PhilosophyUniversity of Oxford Oxford OX2 6GG U.K
  • 7 Philosophie DepartmentCollège André- Laurendeau Montreal QC H8N 2J4 Canada
  • 8 Philosophy DepartmentPace University New York NY 10038 USA
  • 9 Wexner Medical CenterDepartment of RadiologyThe Ohio State University Columbus OH 43210 USA
  • 10 Department of RadiologyHumanitas Research Hospital 20089 Milan Italy
  • 11 Faculty of Science, Agriculture, Business and LawUniversity of New England Armidale NSW 2351 Australia
  • 12 European Centre of Excellence on the Regulation of Robotics & AIScuola Superiore Sant'Anna 56127 Pisa Italy
  • 13 Group of Computer ArchitectureUniversity of Bremen 28359 Bremen Germany
  • 14 Department of MedicineDivision of Infectious Diseases and ImmunologyNew York University Grossman School of Medicine New York NY 10016 USA
  • 15 AI Research SectionDigital Institute 16731 Stockholm Sweden
  • 16 Department of Business Management and AnalyticsArcada University of Applied Sciences 00550 Helsinki Finland
  • 17 Department of Medical and Surgical SpecialtiesRadiological Sciences, and Public HealthUniversity of Brescia 25121 Brescia Italy
  • 18 Ethique GroupeSNCF Reseau SA 93418 La Plaine France
  • 19 Institute of Diagnostic and Interventional Radiology, University Hospital Zurich 8091 Zürich Switzerland
  • 20 Department of Mathematics and Computer ScienceEindhoven University of Technology 5600 MB Eindhoven The Netherlands
  • 21 Data Privacy Advisor, Ericsson 11331 Stockholm Sweden
  • 22 Department of Cardiology4th Gliwice Municipal Hospital 44-100 Gliwice Poland
  • 23 Z-Inspection® Initiative Frankfurt Germany
  • 24 Center for the Study of Ethics in the ProfessionsIllinois Institute of Technology Chicago IL 60616 USA
  • 25 Department of Food and Resource EconomicsUniversity of Copenhagen 1165 Copenhagen Denmark
  • 26 Department of PhysiologyFaculty of MedicineUniversity of Iceland 101 Reykjavik Iceland
  • 27 Cyber Policy CenterStanford University Stanford CA 94305 USA
  • 28 Department for DermatologyHautmedizin Bad Soden 65812 Bad Soden Germany
  • 29 Data Science DepartmentScuola Normale Superiore 56126 Pisa Italy
  • 30 Charité Lab for Artificial Intelligence in MedicineDepartment of Neurosurgery, the QUEST Centre for Responsible ResearchBerlin Institute of Health at Charité, Charité Universitätsmedizin Berlin 10117 Berlin Germany
  • 31 Department of Medical InformaticsErasmus University Medical Center 3015 GD Rotterdam The Netherlands
  • 32 CADS DepartmentSyngenta Research Triangle Park NC 27709 USA
  • 33 Policy ResearchThe Lisbon Council 1040 Brussels Belgium
  • 34 Department of Translational ResearchAcademic RadiologyUniversity of Pisa 56126 Pisa Italy
  • 35 Center for Psychiatry ResearchDepartment of Clinical NeuroscienceKarolinska Institutet 171 77 Stockholm Sweden
  • 36 Department of MedicineThe University of Cambridge, Addenbrooke's Hospital Cambridge CB2 0QQ U.K
  • 37 Department of Computer ScienceUniversity of Pisa 56127 Pisa Italy
  • 38 Division of Medical EthicsDepartment of Population HealthNYU Grossman School of Medicine New York NY 10016 USA
  • 39 School of Healthcare and Social WorkSeinäjoki University of Applied Sciences 60100 Seinäjoki Finland
  • 40 Law DepartmentEuropean University Institute 50139 Firenze Italy
  • 41 Department of Information EngineeringUniversity of Brescia 25121 Brescia Italy
  • 42 Advanced AnalyticsFinnish Tax Administration 00510 Helsinki Finland
  • 43 Artificial Intelligence ResearchAI for Good Foundation El Cerrito CA 94530 USA
  • 44 Data Services Function in Roche Diagnostics, Roche Pune 411005 India
  • 45 Digital SystemsNORCE Norwegian Research Centre AS 5008 Bergen Norway
  • 46 School of ResearchSwinburne University of Technology (Sarawak) Kuching 93350 Malaysia
  • 47 Parallel Computing Labs, Intel Santa Clara CA 95054 USA
  • 48 Department of Psychology and Speech- Language PathologyUniversity of Turku 20500 Turku Finland
  • 49 Department of ValuesTechnology and InnovationDelft University of Technology 2628 BX Delft The Netherlands
  • 50 Computational Vision and Artificial Intelligence LabGoethe University Frankfurt 60325 Frankfurt Germany
  • 51 T. H. Chan School of Public HealthHarvard University Boston MA 02115 USA
IEEE Trans Technol Soc, 2022 Dec;3(4):272-289.
PMID: 36573115 DOI: 10.1109/TTS.2022.3195114

Abstract

This article's main contributions are twofold: 1) to demonstrate how to apply the general European Union's High-Level Expert Group's (EU HLEG) guidelines for trustworthy AI in practice for the domain of healthcare and 2) to investigate the research question of what does "trustworthy AI" mean at the time of the COVID-19 pandemic. To this end, we present the results of a post-hoc self-assessment to evaluate the trustworthiness of an AI system for predicting a multiregional score conveying the degree of lung compromise in COVID-19 patients, developed and verified by an interdisciplinary team with members from academia, public hospitals, and industry in time of pandemic. The AI system aims to help radiologists to estimate and communicate the severity of damage in a patient's lung from Chest X-rays. It has been experimentally deployed in the radiology department of the ASST Spedali Civili clinic in Brescia, Italy, since December 2020 during pandemic time. The methodology we have applied for our post-hoc assessment, called Z-Inspection®, uses sociotechnical scenarios to identify ethical, technical, and domain-specific issues in the use of the AI system in the context of the pandemic.

* Title and MeSH Headings from MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine.

Similar publications